YUNet: Improved YOLOv11 Network for Skyline Detection
Gang Yang, Miao Wang, Quan Zhou, Jiangchuan Li
TL;DR
The paper addresses skyline detection under variable weather and illumination by introducing YUNet, an UNet-like extension of YOLOv11 that uses an encoder–neck–decoder architecture to perform sky segmentation and then extracts the skyline boundary via edge detection. The neck, inspired by PAFPN, fuses multi-scale features to capture large-range contextual information, overcoming limitations of standard UNet and YOLOv11. Empirical results on Skyfinder and CH1 show state-of-the-art performance: IoU for sky segmentation reaches up to $0.9858$ (YUNet-x) and average skyline error is $1.36$ pixels, with PAD defined as $E = \frac{1}{N}\sum_{i=1}^{n}|p_i-p_i^*|$ guiding skyline evaluation. The work demonstrates robust skyline detection suitable for geolocation, flight control, and visual navigation, with a public implementation at the linked repository.
Abstract
Skyline detection plays an important role in geolocalizaion, flight control, visual navigation, port security, etc. The appearance of the sky and non-sky areas are variable, because of different weather or illumination environment, which brings challenges to skyline detection. In this research, we proposed the YUNet algorithm, which improved the YOLOv11 architecture to segment the sky region and extract the skyline in complicated and variable circumstances. To improve the ability of multi-scale and large range contextual feature fusion, the YOLOv11 architecture is extended as an UNet-like architecture, consisting of an encoder, neck and decoder submodule. The encoder extracts the multi-scale features from the given images. The neck makes fusion of these multi-scale features. The decoder applies the fused features to complete the prediction rebuilding. To validate the proposed approach, the YUNet was tested on Skyfinder and CH1 datasets for segmentation and skyline detection respectively. Our test shows that the IoU of YUnet segmentation can reach 0.9858, and the average error of YUnet skyline detection is just 1.36 pixels. The implementation is published at https://github.com/kuazhangxiaoai/SkylineDet-YOLOv11Seg.git.